What's cooking (2010/09 #05)

Junio C Hamano committed Sep 15, 2010 at 15:16 UTC 46b1f9b6e6ce2a4243daab8d1ab5d7bf895f6804
1 file changed +227 -103
whats-cooking.txt
+227 -103
@@ -1,9 +1,9 @@
1 To: git@vger.kernel.org
2 -Subject: What's cooking in git.git (Sep 2010, #04; Sun, 12)
3 -X-master-at: 5879b6bbcaba7696e4abfa717748da166f915405
4 -X-next-at: ee5c7b80a60b21d82897a2c9d9a55bcea89e8c99
2 +Subject: What's cooking in git.git (Sep 2010, #05; Wed, 15)
3 +X-master-at: 7e100effc6c3a8acbb4bf305638c7b324a63236d
4 +X-next-at: bf93f8322c5e01b030ee9a1d3dee19e4d4dbd2d7
5
6 -What's cooking in git.git (Sep 2010, #04; Sun, 12)
6 +What's cooking in git.git (Sep 2010, #05; Wed, 15)
7 --------------------------------------------------
8
9 Here are the topics that have been cooking. Commits prefixed with '-' are
@@ -11,25 +11,110 @@ only in 'pu' while commits prefixed with '+' are in 'next'. The ones
11 marked with '.' do not appear in any of the integration branches, but I am
12 still holding onto them.
13
14 -We have in 'master' is more-or-less the final 1.7.3; it is tagged as rc1.
14 +The tip of 'master' is now at 1.7.3-rc2; I am hoping that we can do the
15 +real 1.7.3 release perhaps this weekend. Please give it a final round of
16 +beating so that we can find and fix the last minute regressions.
17
16 -Please give it a good beating to find regressions.
18 +Thanks.
19
20 --------------------------------------------------
19 -[New Topics]
21 +[Graduated to "master"]
22 +
23 +* ch/filter-branch-deprecate-remap-to-ancestor (2010-08-27) 1 commit
24 + (merged to 'next' on 2010-09-10 at f91d974)
25 + + filter-branch: retire --remap-to-ancestor
26 +
27 +* ch/maint-cannot-create-bundle-error (2010-08-27) 1 commit
28 + (merged to 'next' on 2010-09-10 at ba139f1)
29 + + bundle: detect if bundle file cannot be created
30
31 * dr/maint-ls-tree-prefix-recursion-fix (2010-09-11) 1 commit
32 (merged to 'next' on 2010-09-12 at c9d05a6)
33 + ls-tree $di $dir: do not mistakenly recurse into directories
34
25 -* jn/gitweb-test-lib (2010-09-12) 2 commits
26 - (merged to 'next' on 2010-09-12 at 4621733)
27 - + t/gitweb-lib.sh: Use tabs for indent consistently
28 - + t/gitweb-lib.sh: Use GIT_BUILD_DIR
35 +* jl/fix-test (2010-09-07) 7 commits
36 + (merged to 'next' on 2010-09-10 at 849b1bb)
37 + + t1020: Get rid of 'cd "$HERE"' at the start of each test
38 + + t2016 (checkout -p): add missing &&
39 + + t1302 (core.repositoryversion): style tweaks
40 + + t2105 (gitfile): add missing &&
41 + + t1450 (fsck): remove dangling objects
42 + + tests: subshell indentation stylefix
43 + + Merge branch 'jl/maint-fix-test'
44 + (this branch uses jl/maint-fix-test.)
45 +
46 +* jl/maint-fix-test (2010-09-06) 1 commit
47 + + Several tests: cd inside subshell instead of around
48 + (this branch is used by jl/fix-test.)
49 +
50 +* ks/recursive-rename-add-identical (2010-09-01) 1 commit
51 + (merged to 'next' on 2010-09-03 at 1a5082f)
52 + + RE: [PATCH] Avoid rename/add conflict when contents are identical
53 +
54 +* os/fix-rebase-diff-no-prefix (2010-09-09) 1 commit
55 + (merged to 'next' on 2010-09-10 at f1acd10)
56 + + Add --src/dst-prefix to git-formt-patch in git-rebase.sh
57 +
58 +--------------------------------------------------
59 +[New Topics]
60 +
61 +* ab/send-email-catfile (2010-09-14) 1 commit
62 + - send-email: use catfile() to concatenate files
63 +
64 +Sane; should be safe to be among the first batch after 1.7.3 release.
65 +
66 +* bg/fix-t7003 (2010-09-10) 1 commit
67 + (merged to 'next' on 2010-09-13 at 9d7756e)
68 + + t7003: Use test_commit instead of custom function
69 +
70 +Sane; should be safe to be among the first batch after 1.7.3 release.
71 +
72 +* jc/grep-header-all-match-fix (2010-09-12) 2 commits
73 + - log --author: take union of multiple "author" requests
74 + - grep: move logic to compile header pattern into a separate helper
75 + (this branch is used by jc/grep-header-all-match-fix-debug.)
76 +
77 +We might want to give a more comprehensive revamp to the "filter by
78 +grepping the commit log message" feature some day, somehow allowing the
79 +full "git grep" boolean expression. But until then, this should suffice.
80 +
81 +* jc/grep-header-all-match-fix-debug (2010-09-13) 1 commit
82 + - grep debugging, just in case
83 + (this branch uses jc/grep-header-all-match-fix.)
84 +
85 +* jc/no-branch-name-with-dash-at-front (2010-09-14) 1 commit
86 + - disallow branch names that start with a hyphen
87 +
88 +This also came up at $WORK; should be safe to be among the first batch
89 +after 1.7.3 release.
90 +
91 +* dm/mergetool-vimdiff (2010-09-14) 2 commits
92 + - mergetool-lib: add a three-way diff view for vim/gvim
93 + - mergetool-lib: combine vimdiff and gvimdiff run blocks
94 +
95 +The bottom one should be without controversy. I don't use vimdiff, so
96 +cannot comment on the other one.
97
98 --------------------------------------------------
99 [Stalled]
100
101 +* cb/ignored-paths-are-precious (2010-08-21) 1 commit
102 + - checkout/merge: optionally fail operation when ignored files need to be overwritten
103 +
104 +I fixed up a trivial error in the configuration parser (which didn't
105 +return immediately after parsing), and while at it flipped the default
106 +value of the configuration and the command line option to false.
107 +
108 +This probably needs tests; also we know of longstanding bugs in related
109 +area that needs to be addressed---they do not have to be part of this
110 +series but their reproduction recipe would belong to the test script for
111 +this topic.
112 +
113 +It may hurt users to make the new feature on by default, especially the
114 +ones with subdirectories that come and go. For a patch that changes a
115 +long established semantics, it came a bit late in the cycle, so it will
116 +have to wait until the next release.
117 +
118 * jj/icase-directory (2010-08-16) 6 commits
119 (merged to 'next' on 2010-08-26 at 9d8e1bc)
120 + Support case folding in git fast-import when core.ignorecase=true
@@ -42,28 +127,6 @@ Please give it a good beating to find regressions.
127 Depends on GNU FNM_CASEFOLD. Presumably a bit of tweak in Makefile for
128 non-windows but non-GNU platforms is all it takes?
129
45 -* by/line-log (2010-09-11) 18 commits
46 - (merged to 'next' on 2010-09-12 at e29a029)
47 - + log -L: do not free parents lists we might need again
48 - (merged to 'next' on 2010-08-26 at a160ff1)
49 - + Document line history browser
50 - + Add tests for line history browser
51 - + Add --full-line-diff option
52 - + Add --graph prefix before line history output
53 - + Add parent rewriting to line history browser
54 - + Make graph_next_line external to other part of git
55 - + Make rewrite_parents public to other part of git
56 - + Hook line history into cmd_log, ensuring a topo-ordered walk
57 - + Print the line log
58 - + map/take range to the parent of commits
59 - + Add range clone functions
60 - + Export three functions from diff.c
61 - + Parse the -L options
62 - + Refactor parse_loc
63 - + Add the basic data structure for line level history
64 - + parse-options: add two helper functions
65 - + parse-options: enhance STOP_AT_NON_OPTION
66 -
130 * il/rfc-remote-fd-ext (2010-08-26) 7 commits
131 - remote-ext/fd style fixup
132 - Fix build on Windows
@@ -99,57 +162,90 @@ current usage context.
162 --------------------------------------------------
163 [Cooking]
164
102 -* jl/fix-test (2010-09-07) 7 commits
103 - (merged to 'next' on 2010-09-10 at 849b1bb)
104 - + t1020: Get rid of 'cd "$HERE"' at the start of each test
105 - + t2016 (checkout -p): add missing &&
106 - + t1302 (core.repositoryversion): style tweaks
107 - + t2105 (gitfile): add missing &&
108 - + t1450 (fsck): remove dangling objects
109 - + tests: subshell indentation stylefix
110 - + Merge branch 'jl/maint-fix-test'
111 - (this branch uses jl/maint-fix-test.)
165 +* by/line-log (2010-09-11) 18 commits
166 + (merged to 'next' on 2010-09-12 at e29a029)
167 + + log -L: do not free parents lists we might need again
168 + (merged to 'next' on 2010-08-26 at a160ff1)
169 + + Document line history browser
170 + + Add tests for line history browser
171 + + Add --full-line-diff option
172 + + Add --graph prefix before line history output
173 + + Add parent rewriting to line history browser
174 + + Make graph_next_line external to other part of git
175 + + Make rewrite_parents public to other part of git
176 + + Hook line history into cmd_log, ensuring a topo-ordered walk
177 + + Print the line log
178 + + map/take range to the parent of commits
179 + + Add range clone functions
180 + + Export three functions from diff.c
181 + + Parse the -L options
182 + + Refactor parse_loc
183 + + Add the basic data structure for line level history
184 + + parse-options: add two helper functions
185 + + parse-options: enhance STOP_AT_NON_OPTION
186
113 -* jl/maint-fix-test (2010-09-06) 1 commit
114 - + Several tests: cd inside subshell instead of around
115 - (this branch is used by jl/fix-test.)
187 +How would one trace history of lines 440,450 in t/t7300-clean.sh starting
188 +from the tip of 'pu'? "git log -p -L 440,450 pu -- t/t7300-clean.sh"
189 +would be the most natural but it does not work (it wants "pu" to be a
190 +path). "git log -p -L 440,450 t/t7300-clean.sh pu" doesn't work either.
191 +
192 +The command line parser of this series needs to be rethought (didn't I say
193 +that many times? Perhaps I was ignored). Might want to kick this out of
194 +'next' and move to the stalled category after 1.7.3.
195 +
196 +* jn/gitweb-test-lib (2010-09-12) 2 commits
197 + (merged to 'next' on 2010-09-12 at 4621733)
198 + + t/gitweb-lib.sh: Use tabs for indent consistently
199 + + t/gitweb-lib.sh: Use GIT_BUILD_DIR
200 +
201 +Sane; should be safe to be among the first batch after 1.7.3 release.
202
203 * nd/index-doc (2010-09-06) 1 commit
204 (merged to 'next' on 2010-09-10 at 2ab8572)
205 + doc: technical details about the index file format
206
121 -Half-written but it is a good start.
207 +Half-written but it is a good start. I may need to give some help in
208 +describing more recent index extensions.
209
210 * tr/send-email-refuse-sending-unedited-cover-letter (2009-06-08) 1 commit
211 - send-email: Refuse to send cover-letter template subject
212
213 +Sane; should be safe to be among the first batch after 1.7.3 release.
214 +
215 * bc/fortran-userdiff (2010-09-10) 1 commit
216 (merged to 'next' on 2010-09-10 at b76b831)
217 + userdiff.c: add builtin fortran regex patterns
218
219 +Sane; should be safe to be among the first batch after 1.7.3 release.
220 +
221 * en/rename-d-f (2010-09-08) 2 commits
131 - - merge-recursive: D/F conflicts where was_a_dir/file -> was_a_dir
132 - - t3509: Add rename + D/F conflict testcase that recursive strategy fails
222 + (merged to 'next' on 2010-09-15 at d6a2e1e)
223 + + merge-recursive: D/F conflicts where was_a_dir/file -> was_a_dir
224 + + t3509: Add rename + D/F conflict testcase that recursive strategy fails
225 +
226 +I am not entirely convinced this is a regression free band-aid; need to
227 +look at this a few more times.
228
229 * jk/read-tree-empty (2010-09-10) 1 commit
230 (merged to 'next' on 2010-09-10 at 2f3a995)
231 + read-tree: deprecate syntax without tree-ish args
232
233 +Sane; should be safe to be among the first batch after 1.7.3 release.
234 +
235 * kf/post-receive-sample-hook (2010-09-10) 1 commit
236 (merged to 'next' on 2010-09-10 at 53ebeba)
237 + post-receive-email: ensure sent messages are not empty
238
239 +I notice that it uses "PAGER= generate_email" where generate_email is a
240 +shell function, which may break in some implementations of POSIX /bin/sh.
241 +This is not a regression (the original also had the same issue), but
242 +somebody who cares enough might want to look into it.
243 +
244 * ml/completion-zsh (2010-09-06) 1 commit
245 (merged to 'next' on 2010-09-10 at b509ebc)
246 + completion: make compatible with zsh
247
146 -* os/fix-rebase-diff-no-prefix (2010-09-09) 1 commit
147 - (merged to 'next' on 2010-09-10 at f1acd10)
148 - + Add --src/dst-prefix to git-formt-patch in git-rebase.sh
149 -
150 -* ks/recursive-rename-add-identical (2010-09-01) 1 commit
151 - (merged to 'next' on 2010-09-03 at 1a5082f)
152 - + RE: [PATCH] Avoid rename/add conflict when contents are identical
248 +Comments from bash users regarding regressions?
249
250 * po/sendemail (2010-09-06) 3 commits
251 (merged to 'next' on 2010-09-10 at d39274d)
@@ -157,26 +253,26 @@ Half-written but it is a good start.
253 + Remove @smtp_host_parts variable as not used.
254 + Minor indentation fix.
255
160 -* ch/filter-branch-deprecate-remap-to-ancestor (2010-08-27) 1 commit
161 - (merged to 'next' on 2010-09-10 at f91d974)
162 - + filter-branch: retire --remap-to-ancestor
163 -
164 -* ch/maint-cannot-create-bundle-error (2010-08-27) 1 commit
165 - (merged to 'next' on 2010-09-10 at ba139f1)
166 - + bundle: detect if bundle file cannot be created
256 +Comments from potential users?
257
258 * rr/format-patch-count-without-merges (2010-08-28) 2 commits
259 (merged to 'next' on 2010-09-10 at c23bd48)
260 + format-patch: Don't go over merge commits
261 + t4014-format-patch: Call test_tick before committing
262
263 +Should be safe to be among the first batch after 1.7.3 release.
264 +
265 * as/daemon-multi-listen (2010-08-29) 2 commits
266 (merged to 'next' on 2010-09-10 at 69e4b23)
267 + daemon: allow more than one host address given via --listen
268 + daemon: add helper function setup_named_sock()
269
178 -I think the idea is sound. It came a bit late in the cycle for the next
179 -release and probably needs a bit more polish.
270 +I think the idea is sound. It probably needs a bit more polish.
271 +
272 +If you give two --listen options and got some but not all sockets, the
273 +code does not say anything even when no socket was created for one of
274 +them, which is questionable. Might want to kick this out of 'next' and
275 +move to the stalled category after 1.7.3.
276
277 * jc/pickaxe-grep (2010-08-31) 4 commits
278 (merged to 'next' on 2010-09-10 at 3604fc3)
@@ -188,27 +284,31 @@ release and probably needs a bit more polish.
284 This is a re-roll of "grepping inside the log -p output" which is a
285 feature that is often asked for when people hear about -S option.
286
287 +Should be safe to be among the first batch after 1.7.3 release.
288 +
289 * jl/fetch-submodule-recursive (2010-08-30) 2 commits
290 (merged to 'next' on 2010-09-10 at 67a5138)
291 + Submodules: Add the new "fetch" config option for fetch and pull
292 + fetch/pull: Recursively fetch populated submodules
293
196 -I think the idea is sound and hopefully the end result will help the heavy
197 -users of subprojects. It came a bit late in the cycle for the next
198 -release, though.
294 +Further work expected after 1.7.3 between Jens and Kevin.
295
296 * en/object-list-with-pathspec (2010-08-26) 2 commits
297 - Make rev-list --objects work together with pathspecs
298 - Add testcases showing how pathspecs are ignored with rev-list --objects
299
300 +There is a reroll helped by Duy, which I expect to be rerolled after
301 +en/tree-walk-optim hits master.
302 +
303 * jf/merge-ignore-ws (2010-08-26) 4 commits
205 - - merge-recursive: options to ignore whitespace changes
206 - - merge-recursive --patience
207 - - ll-merge: replace flag argument with options struct
208 - - merge-recursive: expose merge options for builtin merge
304 + (merged to 'next' on 2010-09-15 at bf93f83)
305 + + merge-recursive: options to ignore whitespace changes
306 + + merge-recursive --patience
307 + + ll-merge: replace flag argument with options struct
308 + + merge-recursive: expose merge options for builtin merge
309
210 -I think the idea is sound. It came a bit late in the cycle for the next
211 -release, though.
310 +Possibly one of the star features of the release after 1.7.3, whether it
311 +is called 1.7.4 or 1.8.0.
312
313 * rr/fmt-merge-msg (2010-09-08) 5 commits
314 (merged to 'next' on 2010-09-10 at 40fe79e)
@@ -218,6 +318,8 @@ release, though.
318 + merge: Make '--log' an integer option for number of shortlog entries
319 + fmt_merge_msg: Change fmt_merge_msg API to accept shortlog_len
320
321 +Should be safe to be among the first batch after 1.7.3 release.
322 +
323 * tr/merge-unborn-clobber (2010-08-22) 1 commit
324 - Exhibit merge bug that clobbers index&WT
325
@@ -228,32 +330,14 @@ release, though.
330 + tree-walk: Correct bitrotted comment about tree_entry()
331 + Document pre-condition for tree_entry_interesting
332
231 -I think the idea is sound. For a patch that touches fairly core part of
232 -the system, it came a bit late in the cycle, though.
233 -
234 -* cb/ignored-paths-are-precious (2010-08-21) 1 commit
235 - - checkout/merge: optionally fail operation when ignored files need to be overwritten
236 -
237 -I fixed up a trivial error in the configuration parser (which didn't
238 -return immediately after parsing), and while at it flipped the default
239 -value of the configuration and the command line option to false.
240 -
241 -This probably needs tests; also we know of longstanding bugs in related
242 -area that needs to be addressed---they do not have to be part of this
243 -series but their reproduction recipe would belong to the test script for
244 -this topic.
245 -
246 -It may hurt users to make the new feature on by default, especially the
247 -ones with subdirectories that come and go. For a patch that changes a
248 -long established semantics, it came a bit late in the cycle, so it will
249 -have to wait until the next release.
333 +Need to look at this a few more times to convince myself that this is Ok;
334 +will happen after 1.7.3.
335
336 * po/etc-gitattributes (2010-09-01) 1 commit
337 (merged to 'next' on 2010-09-10 at c503fd5)
338 + Add global and system-wide gitattributes
339
255 -I think the idea is sound. For a patch that touches fairly core part of
256 -the system, it came a bit late in the cycle, though.
340 +Sane; should be safe to be among the first batch after 1.7.3 release.
341
342 * gb/shell-ext (2010-08-27) 6 commits
343 (merged to 'next' on 2010-08-27 at e1ef3c1)
@@ -265,11 +349,49 @@ the system, it came a bit late in the cycle, though.
349 + Add interactive mode to git-shell for user-friendliness
350 + Allow creation of arbitrary git-shell commands
351
268 -I am not very happy about adding these backdoors to git-shell, which is
269 -primarily a security mechanism, and obviously security and backdoor do not
270 -mix well.
271 -
272 -* ab/i18n (2010-09-10) 98 commits
352 +Should be safe to be among the first batch after 1.7.3 release.
353 +
354 +* ab/i18n (2010-09-14) 140 commits
355 + - gettextize: git-bisect bisect_next_check "You need to" message
356 + - gettextize: git-bisect [Y/n] messages
357 + - gettextize: git-bisect bisect_replay + $1 messages
358 + - gettextize: git-bisect bisect_reset + $1 messages
359 + - gettextize: git-bisect bisect_run + $@ messages
360 + - gettextize: git-bisect die + eval_gettext messages
361 + - gettextize: git-bisect die + gettext messages
362 + - gettextize: git-bisect echo + eval_gettext message
363 + - gettextize: git-bisect echo + gettext messages
364 + - gettextize: git-bisect gettext + echo message
365 + - gettextize: git-bisect add git-sh-i18n
366 + - gettextize: git-stash drop_stash say/die messages
367 + - gettextize: git-stash "unknown option" message
368 + - gettextize: git-stash die + eval_gettext $1 messages
369 + - gettextize: git-stash die + eval_gettext $* messages
370 + - gettextize: git-stash die + eval_gettext messages
371 + - gettextize: git-stash die + gettext messages
372 + - gettextize: git-stash say + gettext messages
373 + - gettextize: git-stash echo + gettext message
374 + - gettextize: git-stash add git-sh-i18n
375 + - gettextize: git-submodule "blob" and "submodule" messages
376 + - gettextize: git-submodule "path not initialized" message
377 + - gettextize: git-submodule "[...] path is ignored" message
378 + - gettextize: git-submodule "Entering [...]" message
379 + - gettextize: git-submodule $errmsg messages
380 + - gettextize: git-submodule "Submodule change[...]" messages
381 + - gettextize: git-submodule "cached cannot be used" message
382 + - gettextize: git-submodule $update_module say + die messages
383 + - gettextize: git-submodule die + eval_gettext messages
384 + - gettextize: git-submodule say + eval_gettext messages
385 + - gettextize: git-submodule echo + eval_gettext messages
386 + - gettextize: git-submodule add git-sh-i18n
387 + - gettextize: git-pull "rebase against" / "merge with" messages
388 + - gettextize: git-pull "[...] not currently on a branch" message
389 + - gettextize: git-pull "You asked to pull" message
390 + - gettextize: git-pull split up "no candidate" message
391 + - gettextize: git-pull eval_gettext + warning message
392 + - gettextize: git-pull eval_gettext + die message
393 + - gettextize: git-pull die messages
394 + - gettextize: git-pull add git-sh-i18n
395 - gettext docs: add "Testing marked strings" section to po/README
396 - gettext docs: the Git::I18N Perl interface
397 - gettext docs: the git-sh-i18n.sh Shell interface
@@ -281,6 +403,7 @@ mix well.
403 - po/en_GB.po: add British English translation
404 - po/de.po: add German translation
405 - po/is.po: add Icelandic translation
406 + - Makefile: only add gettext tests on XGETTEXT_INCLUDE_TESTS=YesPlease
407 - gettext docs: add po/README file documenting Git's gettext
408 - gettextize: git-am printf(1) message to eval_gettext
409 - gettextize: git-am core say messages
@@ -363,8 +486,9 @@ mix well.
486 - Makefile: tell xgettext(1) that our source is in UTF-8
487 - Makefile: provide a --msgid-bugs-address to xgettext(1)
488 - Makefile: A variable for options used by xgettext(1) calls
366 - - tests: locate i18n lib&data correctly under --valgrind
489 + - gettext tests: locate i18n lib&data correctly under --valgrind
490 - gettext: setlocale(LC_CTYPE, "") breaks Git's C function assumptions
368 - - tests: rename test to work around GNU gettext bug
369 - - Add infrastructure for translating Git with gettext
491 + - gettext tests: rename test to work around GNU gettext bug
492 + - gettext: add infrastructure for translating Git with gettext
493 - builtin: use builtin.h for all builtin commands
494 + - tests: use test_cmp instead of piping to diff(1)